Luis Carlos,

> The similarity is based on 30(could be more) different technical
> construction properties like
> * Material (7 classification)
> * Hole type (3 different holes)
> * shoulder type (3)
> * chamfer type
> * insulation type
> * terminal type
> * shape

30 Properties is not a problem...

> So, when comparing brush 1 with brush2, the first one can have a shape

and

> the second one a different shape. So the similarity will be zero.

There is

> a set of pre-defined values that must match in order to have a minimum
> level of similarity (shape,application, both with or without shunt)

What is your estimate of the largest/smallest number of brushes which
share the same minimum similarity?

The fact that there are some required common point will significantly
reduce the actual number of brushes that need to be evaluated to find
the similarity.

> After this minimal set is reached, others comparisons will be

performed

> and the result will be a percentage stored in column similarity at the
> table described.

Please provide an outline of this "similarity formula", it will help in
understanding the nature of the comparisons.


Separately, I'm not clear on the analysis to be performed.

Are you looking to take a brush and then find all similar brushes? Or

Are you looking for the set of brushes which share the same level of
similarity?


Sean